Likelihood-based estimation of the regression model with scrambled responses

A significant problem in the collection of responses to potentially sensitive questions, such as relating to illegal, immoral or embarrassing activities, is non-sampling error due to refusal to respond or false responses. Eichhorn & Hayre (1983) suggested the use of scrambled responses to reduce this form of bias. This paper considers a linear regression model in which the dependent variable is unobserved but for which the sum or product with a scrambling random variable of known distribution, is known. The performance of two likelihood-based estimators is investigated, namely of a Bayesian estimator achieved through a Markov chain Monte Carlo (MCMC) sampling scheme, and a classical maximum-likelihood estimator. These two estimators and an estimator suggested by Singh, Joarder & King (1996) are compared. Monte Carlo results show that the Bayesian estimator outperforms the classical estimators in almost all cases, and the relative performance of the Bayesian estimator improves as the responses become more scrambled.

The marginal dentition of the Australian lungfish, Neoceratodus forsteri (Osteichthyes : Dipnoi)

The dentary, a component of the transient marginal dentition found in the mandible of juveniles of the living Australian lungfish Neoceratodus forsteri, is a tooth plate exactly comparable to the tooth plates with radiating ridges that make up the marginal dentitions of Devonian dipnoans like Andreyevichthys, Orlovichthys and Ichnomylax. In N. forsteri, the dentary consists of two ridges, set almost in line with each other, and growing by the addition of cusps, of increasing sizes, to the extremity of each ridge. It is therefore equivalent to two ridges of a more normal tooth plate with radiating ridges. Despite its appearance, as a long row of sharp cusps ankylosed to a slender bone, and its position, embedded in soft tissue above the anterolabial margin of Meckel's cartilage, it is a tooth plate and is not comparable to the marginal dentitions of other vertebrates. Structure and development of the transient marginal dentition of this lungfish is another indication that dipnoans may not be the sister group of tetrapods.

Maximum Likelihood Estimation of Mixture Densities for Binned and Truncated Multivariate Data

Binning and truncation of data are common in data analysis and machine learning. This paper addresses the problem of fitting mixture densities to multivariate binned and truncated data. The EM approach proposed by McLachlan and Jones (Biometrics, 44: 2, 571-578, 1988) for the univariate case is generalized to multivariate measurements. The multivariate solution requires the evaluation of multidimensional integrals over each bin at each iteration of the EM procedure. Naive implementation of the procedure can lead to computationally inefficient results. To reduce the computational cost a number of straightforward numerical techniques are proposed. Results on simulated data indicate that the proposed methods can achieve significant computational gains with no loss in the accuracy of the final parameter estimates. Furthermore, experimental results suggest that with a sufficient number of bins and data points it is possible to estimate the true underlying density almost as well as if the data were not binned. The paper concludes with a brief description of an application of this approach to diagnosis of iron deficiency anemia, in the context of binned and truncated bivariate measurements of volume and hemoglobin concentration from an individual's red blood cells.

Deviance and marginal occupations: the case of taxi drivers

The taxi industry provides a strategic site to explore workplace deviance in low supervision, low status, occupational settings. Despite this theoretical opportunity and the objective importance of the taxi industry worldwide, very little is known about deviance among taxi drivers. Making use of interview data, this exploratory study maps out forms of workplace deviance and the explanations given for them by a sample of male taxi drivers. Major illegal activities reported included speeding, driving unsafe vehicles, taking drugs, and fraud. Theories pertaining to worker alienation, stress management, victim precipitation, and social control were relevant to the rationalizations some of the drivers provided to account for their illegal behaviors. We conclude that the occupational culture of taxi drivers and the structure of the taxi industry facilitate the forms of deviance reported here.

Rock fracture mean trace length estimation and confidence interval calculation using maximum likelihood methods

There has been a resurgence of interest in the mean trace length estimator of Pahl for window sampling of traces. The estimator has been dealt with by Mauldon and Zhang and Einstein in recent publications. The estimator is a very useful one in that it is non-parametric. However, despite some discussion regarding the statistical distribution of the estimator, none of the recent works or the original work by Pahl provide a rigorous basis for the determination a confidence interval for the estimator or a confidence region for the estimator and the corresponding estimator of trace spatial intensity in the sampling window. This paper shows, by consideration of a simplified version of the problem but without loss of generality, that the estimator is in fact the maximum likelihood estimator (MLE) and that it can be considered essentially unbiased. As the MLE, it possesses the least variance of all estimators and confidence intervals or regions should therefore be available through application of classical ML theory. It is shown that valid confidence intervals can in fact be determined. The results of the work and the calculations of the confidence intervals are illustrated by example. (C) 2003 Elsevier Science Ltd. All rights reserved.

Clinical and economic impact of exercise electrocardiography and exercise echocardiography in clinical practice

Background Patients with known or suspected coronary disease are often investigated to facilitate risk assessment. We sought to examine the cost-effectiveness of strategies based on exercise echocardiography and exercise electrocardiography. Methods and results We studied 7656 patients undergoing exercise testing; of whom half underwent exercise echocardiography. Risk was defined with the Duke treadmill score for those undergoing exercise electrocardiography alone, and by the extent of ischaemia by exercise echocardiography. Cox proportional hazards models, risk adjusted for pretest likelihood of coronary artery disease, were used to estimate time to cardiac death or myocardial infarction. Costs (including diagnostic and revascularisation procedures, hospitalisations, and events) were calculated, inflation-corrected to year 2000 using Medicare trust fund rates and discounted at a rate of 5%. A decision model was employed to assess the marginal cost effectiveness (cost/life year saved) of exercise echo compared with exercise electrocardiography. Exercise echocardiography identified more patients as low-risk (51% vs 24%, p<0.001), and fewer as intermediate- (27% vs 51%, p<0.001) and high-risk (22% vs 4%); survival was greater in low- and intermediate- risk and less in high-risk patients. Although initial procedural costs and revascularisation costs (in intermediate- high risk patients) were greater, exercise echocardiography was associated with a greater incremental life expectancy (0.2 years) and a lower use of additional diagnostic procedures when compared with exercise electrocardiography (especially in lower risk patients). Using decision analysis, exercise echocardiography (Euro 2615/life year saved) was more cost effective than exercise electrocardiography. Conclusion Exercise echocardiography may enhance cost-effectiveness for the detection and management of at risk patients with known or suspected coronary disease. (C) 2003 Published by Elsevier Science Ltd on behalf of The European Society of Cardiology.

Partitioned likelihood support and the evaluation of data set conflict

In simultaneous analyses of multiple data partitions, the trees relevant when measuring support for a clade are the optimal tree, and the best tree lacking the clade (i.e., the most reasonable alternative). The parsimony-based method of partitioned branch support (PBS) forces each data set to arbitrate between the two relevant trees. This value is the amount each data set contributes to clade support in the combined analysis, and can be very different to support apparent in separate analyses. The approach used in PBS can also be employed in likelihood: a simultaneous analysis of all data retrieves the maximum likelihood tree, and the best tree without the clade of interest is also found. Each data set is fitted to the two trees and the log-likelihood difference calculated, giving partitioned likelihood support (PLS) for each data set. These calculations can be performed regardless of the complexity of the ML model adopted. The significance of PLS can be evaluated using a variety of resampling methods, such as the Kishino-Hasegawa test, the Shimodiara-Hasegawa test, or likelihood weights, although the appropriateness and assumptions of these tests remains debated.

The transform likelihood ratio method for rare event simulation with heavy tails

We present a novel method, called the transform likelihood ratio (TLR) method, for estimation of rare event probabilities with heavy-tailed distributions. Via a simple transformation ( change of variables) technique the TLR method reduces the original rare event probability estimation with heavy tail distributions to an equivalent one with light tail distributions. Once this transformation has been established we estimate the rare event probability via importance sampling, using the classical exponential change of measure or the standard likelihood ratio change of measure. In the latter case the importance sampling distribution is chosen from the same parametric family as the transformed distribution. We estimate the optimal parameter vector of the importance sampling distribution using the cross-entropy method. We prove the polynomial complexity of the TLR method for certain heavy-tailed models and demonstrate numerically its high efficiency for various heavy-tailed models previously thought to be intractable. We also show that the TLR method can be viewed as a universal tool in the sense that not only it provides a unified view for heavy-tailed simulation but also can be efficiently used in simulation with light-tailed distributions. We present extensive simulation results which support the efficiency of the TLR method.

Estimating disease likelihood: a case of rubbery figures

In diagnosis and prognosis, we should avoid intuitive “guesstimates” and seek a validated numerical aid

Modelling cell lifespan and proliferation: is likelihood to die or to divide independent of age?

In cell lifespan studies the exponential nature of cell survival curves is often interpreted as showing the rate of death is independent of the age of the cells within the population. Here we present an alternative model where cells that die are replaced and the age and lifespan of the population pool is monitored until a, steady state is reached. In our model newly generated individual cells are given a determined lifespan drawn from a number of known distributions including the lognormal, which is frequently found in nature. For lognormal lifespans the analytic steady-state survival curve obtained can be well-fit by a single or double exponential, depending on the mean and standard deviation. Thus, experimental evidence for exponential lifespans of one and/or two populations cannot be taken as definitive evidence for time and age independence of cell survival. A related model for a dividing population in steady state is also developed. We propose that the common adoption of age-independent, constant rates of change in biological modelling may be responsible for significant errors, both of interpretation and of mathematical deduction. We suggest that additional mathematical and experimental methods must be used to resolve the relationship between time and behavioural changes by cells that are predominantly unsynchronized.

Higher-order CpG-DNA stimulation reveals distinct activation requirements for marginal zone and follicular B cells in lupus mice

Mouse follicular B cells express TLR9 and respond vigorously to stimulation with single-stranded CpG-oligodeoxynucleotides (ODN). Surprisingly, follicular B cells do not respond to direct stimulation with other TLR9 ligands, such as bacterial DNA or class A(D) CpG-ODN capable of forming higher-order structures, unless other cell types are present. Here, we show that priming with interferons or with B cell-activating factor, or simultaneous co-engagement of the B cell receptor for antigen (BCR), can overcome this unresponsiveness. The effect of interferons occurs at the transcriptional level and is mediated through an autocrine/paracrine loop, which is dependent on IRF-1, IL-6 and IL-12 p40. We hypothesize that the lack of bystander activation of follicular B cells with more complex CpG ligands may be an important safety mechanism for avoiding autoimmunity. This will prevent resting B cells from responding to foreign or self-derived hypomethylated double-stranded CpG ligands unless these ligands are either delivered through the B cell receptor or under conditions where B cells are simultaneously co-engaged by activated plasmacytoid dendritic cells or TH1 cells. A corollary is that the heightened responsiveness of lupus B cells to TLR9-induced stimulation cannot be ascribed to unprimed follicular B cells, but is rather mediated by hypersensitive marginal zone B cells.

Fracture load and marginal fit of shrinkage-free ZrSiO4 all-ceramic crowns after chewing simulation

The aim of this in vitro study was to evaluate the fracture load and marginal accuracy of crowns made from a shrinkage-free ZrSiO4 ceramic cemented with glass-ionomer or composite cement after chewing simulation. Thirty-two human mandibular molars were randomly divided into two groups. All teeth were prepared for and restored with shrinkage-free ZrSiO4 ceramic crowns (Everest HPC (R), KaVo). The crowns of group A (N = 16) were luted to the teeth using KetacCem (R) and group B (N = 16) were adhesively cemented using Panavia (R) 21EX. Measurements of the marginal accuracy before and after cementation were made using replicas and an image analysis system. All specimens were exposed to 1.2 million cycles of thermo-mechanical fatigue in a chewing simulator. Surviving specimens were subsequently loaded until fracture in a static testing device. Fracture loads (N) were recorded. All specimens survived chewing simulation. The mean fracture loads (+/- s.d.) were Group A, 1622 N (+/- 433); group B, 1957 N (+/- 806). There was no significant difference between the two groups (P > 0.05). The marginal gap values before cementation were (mean +/- s.d.): Group A, 32.7 mu m (+/- 6.8); group B, 33.0 mu m (+/- 6.7).The mean marginal gap values after cementation were (+/- s.d.): Group A, 44.6 mu m (+/- 6.7); group B, 46.6 mu m (+/- 7.7). The marginal openings were significantly higher after cementation for both groups (P < 0.05). All test groups demonstrated fracture load and marginal accuracy values within the range of clinical acceptability.